Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userAn Improved Framework of GPU Computing for CFD Applications on Structured Grids using OpenACC
157
0
0.0
(
0
)
Added by
Weicheng Xue
Publication date
2020
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
This paper is focused on improving multi-GPU performance of a research CFD code on structured grids. MPI and OpenACC directives are used to scale the code up to 16 GPUs. This paper shows that using 16 P100 GPUs and 16 V100 GPUs can be 30$times$ and 70$times$ faster than 16 Xeon CPU E5-2680v4 cores for three different test cases, respectively. A series of performance issues related to the scaling for the multi-block CFD code are addressed by applying various optimizations. Performance optimizations such as the pack/unpack message method, removing temporary arrays as arguments to procedure calls, allocating global memory for limiters and connected boundary data, reordering non-blocking MPI I_send/I_recv and Wait calls, reducing unnecessary implicit derived type member data movement between the host and the device and the use of GPUDirect can improve the compute utilization, memory throughput, and asynchronous progression in the multi-block CFD code using modern programming features.
rate research
Read More
This paper investigates the multi-GPU performance of a 3D buoyancy driven cavity solver using MPI and OpenACC directives on different platforms. The paper shows that decomposing the total problem in different dimensions affects the strong scaling performance significantly for the GPU. Without proper performance optimizations, it is shown that 1D domain decomposition scales poorly on multiple GPUs due to the noncontiguous memory access. The performance using whatever decompositions can be benefited from a series of performance optimizations in the paper. Since the buoyancy driven cavity code is latency-bounded on the clusters examined, a series of optimizations both agnostic and tailored to the platforms are designed to reduce the latency cost and improve memory throughput between hosts and devices efficiently. First, the parallel message packing/unpacking strategy developed for noncontiguous data movement between hosts and devices improves the overall performance by about a factor of 2. Second, transferring different data based on the stencil sizes for different variables further reduces the communication overhead. These two optimizations are general enough to be beneficial to stencil computations having ghost changes on all of the clusters tested. Third, GPUDirect is used to improve the communication on clusters which have the hardware and software support for direct communication between GPUs without staging CPUs memory. Finally, overlapping the communication and computations is shown to be not efficient on multi-GPUs if only using MPI or MPI+OpenACC. Although we believe our implementation has revealed enough overlap, the actual running does not utilize the overlap well due to a lack of asynchronous progression.
We present new results on the strong parallel scaling for the OpenACC-accelerated implementation of the high-order spectral element fluid dynamics solver Nek5000. The test case considered consists of a direct numerical simulation of fully-developed turbulent flow in a straight pipe, at two different Reynolds numbers $Re_tau=360$ and $Re_tau=550$, based on friction velocity and pipe radius. The strong scaling is tested on several GPU-enabled HPC systems, including the Swiss Piz Daint system, TACCs Longhorn, Julichs JUWELS Booster, and Berzelius in Sweden. The performance results show that speed-up between 3-5 can be achieved using the GPU accelerated version compared with the CPU version on these different systems. The run-time for 20 timesteps reduces from 43.5 to 13.2 seconds with increasing the number of GPUs from 64 to 512 for $Re_tau=550$ case on JUWELS Booster system. This illustrates the GPU accelerated version the potential for high throughput. At the same time, the strong scaling limit is significantly larger for GPUs, at about $2000-5000$ elements per rank; compared to about $50-100$ for a CPU-rank.
To accelerate the solution of large eigenvalue problems arising from many-body calculations in nuclear physics on distributed-memory parallel systems equipped with general-purpose Graphic Processing Units (GPUs), we modified a previously developed hybrid MPI/OpenMP implementation of an eigensolver written in FORTRAN 90 by using an OpenACC directives based programming model. Such an approach requires making minimal changes to the original code and enables a smooth migration of large-scale nuclear structure simulations from a distributed-memory many-core CPU system to a distributed GPU system. However, in order to make the OpenACC based eigensolver run efficiently on GPUs, we need to take into account the architectural differences between a many-core CPU and a GPU device. Consequently, the optimal way to insert OpenACC directives may be different from the original way of inserting OpenMP directives. We point out these differences in the implementation of sparse matrix-matrix multiplications (SpMM), which constitutes the main cost of the eigensolver, as well as other differences in the preconditioning step and dense linear algebra operations. We compare the performance of the OpenACC based implementation executed on multiple GPUs with the performance on distributed-memory many-core CPUs, and demonstrate significant speedup achieved on GPUs compared to the on-node performance of a many-core CPU. We also show that the overall performance improvement of the eigensolver on multiple GPUs is more modest due to the communication overhead among different MPI ranks.
To harness the potential of advanced computing technologies, efficient (real time) analysis of large amounts of data is as essential as are front-line simulations. In order to optimise this process, experts need to be supported by appropriate tools that allow to interactively guide both the computation and data exploration of the underlying simulation code. The main challenge is to seamlessly feed the user requirements back into the simulation. State-of-the-art attempts to achieve this, have resulted in the insertion of so-called check- and break-points at fixed places in the code. Depending on the size of the problem, this can still compromise the benefits of such an attempt, thus, preventing the experience of real interactive computing. To leverage the concept for a broader scope of applications, it is essential that a user receives an immediate response from the simulation to his or her changes. Our generic integration framework, targeted to the needs of the computational engineering domain, supports distributed computations as well as on-the-fly visualisation in order to reduce latency and enable a high degree of interactivity with only minor code modifications. Namely, the regular course of the simulation coupled to our framework is interrupted in small, cyclic intervals followed by a check for updates. When new data is received, the simulation restarts automatically with the updated settings (boundary conditions, simulation parameters, etc.). To obtain rapid, albeit approximate feedback from the simulation in case of perpetual user interaction, a multi-hierarchical approach is advantageous. Within several different engineering test cases, we will demonstrate the flexibility and the effectiveness of our approach.
High fidelity Computational Fluid Dynamics simulations are generally associated with large computing requirements, which are progressively acute with each new generation of supercomputers. However, significant research efforts are required to unlock the computing power of leading-edge systems, currently referred to as pre-Exascale systems, based on increasingly complex architectures. In this paper, we present the approach implemented in the computational mechanics code Alya. We describe in detail the parallelization strategy implemented to fully exploit the different levels of parallelism, together with a novel co-execution method for the efficient utilization of heterogeneous CPU/GPU architectures. The latter is based on a multi-code co-execution approach with a dynamic load balancing mechanism. The assessment of the performance of all the proposed strategies has been carried out for airplane simulations on the POWER9 architecture accelerated with NVIDIA Volta V100 GPUs.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
